Re: (2WCCS) Rufen Sie für Abstract

2013-02-11 Thread Alec Teal
We're getting a lot of crap ATM? Does an admin know?

genattrtab regression: infinite loop

2013-02-11 Thread Paulo Matos
Hello, I just noticed when moving port from 4.7.2 to HEAD that genattrtab gets into an infinite loop with: (define_insn "br_pred" [(set (pc) (if_then_else (match_operator 2 "easy_comparison" [(match_operand:BI 0 "register_operand" "c") (const_int 0)])

Re: genattrtab regression: infinite loop

2013-02-11 Thread Richard Biener
On Mon, Feb 11, 2013 at 11:29 AM, Paulo Matos wrote: > Hello, > > I just noticed when moving port from 4.7.2 to HEAD that genattrtab gets into > an infinite loop with: > (define_insn "br_pred" > [(set (pc) > (if_then_else > (match_operator 2 "easy_comparison" > [(matc

RE: genattrtab regression: infinite loop

2013-02-11 Thread Paulo Matos
> -Original Message- > From: Richard Biener [mailto:richard.guent...@gmail.com] > Sent: 11 February 2013 10:56 > To: Paulo Matos > Cc: gcc@gcc.gnu.org > Subject: Re: genattrtab regression: infinite loop > > > Does it work with -fno-aggressive-loop-optimizations? > That's new in 4.8, rig

Re: genattrtab regression: infinite loop

2013-02-11 Thread Richard Biener
On Mon, Feb 11, 2013 at 1:36 PM, Paulo Matos wrote: >> -Original Message- >> From: Richard Biener [mailto:richard.guent...@gmail.com] >> Sent: 11 February 2013 10:56 >> To: Paulo Matos >> Cc: gcc@gcc.gnu.org >> Subject: Re: genattrtab regression: infinite loop >> >> >> Does it work with -f

RE: genattrtab regression: infinite loop

2013-02-11 Thread Paulo Matos
The following self-contained md file generates an infinite loop in genattrtab: (define_attr "fpsize" "short,long" (const_string "short")) (define_attr "fplength" "" (if_then_else (eq_attr "fpsize" "long") (const_int 8) (const_int 4))) (define_insn "br_pred" [(set

Re: Global Value Numbering and dependence on SSA in GCC

2013-02-11 Thread Michael Matz
Hi, On Sun, 10 Feb 2013, Richard Biener wrote: > >> Consider the following example: > >> > >> if (...) > >>a_1 = 5; > >> else if (...) > >>a_2 = 2; > >> else > >>a_3 = 13; > >> > >> # a_4 = PHI > >> return a_4; > >> > >> Do you mean that I treat a

RE: genattrtab regression: infinite loop

2013-02-11 Thread Michael Matz
Hi, On Mon, 11 Feb 2013, Paulo Matos wrote: > The following self-contained md file generates an infinite loop in genattrtab: > (define_attr "fpsize" "short,long" (const_string "short")) > > (define_attr "fplength" "" > (if_then_else > (eq_attr "fpsize" "long") > (const_int

RE: genattrtab regression: infinite loop

2013-02-11 Thread Paulo Matos
> -Original Message- > From: Michael Matz [mailto:m...@suse.de] > Sent: 11 February 2013 15:25 > To: Paulo Matos > Cc: Richard Biener; gcc@gcc.gnu.org > Subject: RE: genattrtab regression: infinite loop > > Only by giving a nicer error message than an infinite loop. Attributes > may not

Possible NULL pointer dereference in GCC's vec code?

2013-02-11 Thread Jeff Law
Consider this code in va_gc::reserve: template void va_gc::reserve (vec *&v, unsigned reserve, bool exact MEM_STAT_DECL) { unsigned alloc = vec_prefix::calculate_allocation (v ? &v->vecpfx_ : 0, reserve, exact); if (!alloc) { ::ggc_free (v); v = NULL;

MIPS predefined macro question (_MIPS_ISA)

2013-02-11 Thread Steve Ellcey
I was looking at a bug in my new memcpy in glibc & newlib and traced it down to this code: #if (_MIPS_ISA == _MIPS_ISA_MIPS4) || (_MIPS_ISA == _MIPS_ISA_MIPS5) || \ (_MIPS_ISA == _MIPS_ISA_MIPS32) || (_MIPS_ISA == _MIPS_ISA_MIPS64) The problem is that GCC is defining _MIPS_ISA to something (

Re: MIPS predefined macro question (_MIPS_ISA)

2013-02-11 Thread Richard Sandiford
Steve Ellcey writes: > I was looking at a bug in my new memcpy in glibc & newlib and traced it > down to this code: > > #if (_MIPS_ISA == _MIPS_ISA_MIPS4) || (_MIPS_ISA == _MIPS_ISA_MIPS5) || \ > (_MIPS_ISA == _MIPS_ISA_MIPS32) || (_MIPS_ISA == _MIPS_ISA_MIPS64) > > The problem is that GCC is

Re: Possible NULL pointer dereference in GCC's vec code?

2013-02-11 Thread Ian Lance Taylor
On Mon, Feb 11, 2013 at 11:48 AM, Jeff Law wrote: > > Consider this code in va_gc::reserve: > > > template > void > va_gc::reserve (vec *&v, unsigned reserve, bool exact > MEM_STAT_DECL) > { > unsigned alloc > = vec_prefix::calculate_allocation (v ? &v->vecpfx_ : 0, reserve,

Re: Possible NULL pointer dereference in GCC's vec code?

2013-02-11 Thread Jeff Law
On 02/11/13 16:46, Ian Lance Taylor wrote: But those cases are passing 1 to vec_safe_reserve, which in turn passes 1 to calculate_allocation, so calculate_allocation can not return 0, so vec_reserve can not return NULL. At least that is how it seems to me. Yes, you're right. I got too deep int